ScotGrid Technical Meeting
Other Institutes
-
Gareth Roy (chair)
David Crooks
Andrew Washbrook
Oliver Smith
Sam Skipsey (minutes)
Gang Qin
-
Durham
-
Gareth noted that we located the problem with GLA ARC-CE on the basis
of Oliver's comments (it was a publishing issue).
Oliver: still have the EMI3 migration ticket - the SE is now migrated,
the only thing left is the CREAM CE, which will simply be
decommissioned after the ARC CE is fully tested. (~weeks)
Gareth asked if Ewan talked to ATLAS about test jobs/queues on the ARC
CE (this is still pending).
-
ECDF
-
AW: Upgrading to EMI3 ongoing. A little frustrating - the two services
pending (deployed a new CREAM CE, and were having pool account
mappings, resolved via NFS issues. Now ready to go, along with the
ARGUS server (resolved via help from DC and Iain Collier - the final
issue was that the ARGUS was talking to the CERN central banning, and
timing out, pointing it at the UK redirector was better). Also a new
BDII ready.
Put these all out and expected a sudden update and refresh but NAGIOS
didn't update. The issue seems to be that the *old* site BDII is being
looked at by (at least some) Top BDIIs (even though GOCDB has been
updated to reflect the new config).
There was some discussion over how this has broken (the suggestion
was to ticket RAL or involve Stephen Burke).
Wahid has been working on bringing the new storage online (and is
planning to install directly to DPM 1.8.8 packages).
There was some discussion of how critical ARGUS Central Banning is to
the overall strategy.
-
Glasgow
-
DC: Tickets are as before - EMI3 (completed - APEL redeployment will
happen after this meeting).
Generally stable, other than ATLAS-job-brokering dearths.
Had comments back from Ewan re the website; now ready to pass the beta
site to DB, for comments. A note, the website is now updated from a
central (BitBucket hosted) repo.
People interested in monitoring may be interested in Monitorama, in
Portland, Oregon (currently ongoing).
GQ: The ARC CE currently passes UK Nagios tests. (The next step is to
deploy runtime env scripts on the WNs, as we don't have shared FS).
Next thing is to deploy a test Condor for investigation.
SS: Next storage plans (possibly HDFS/other block-distributed FS
backends). DPM Collaboration Meeting (they liked Wahid and I a lot).
-
AOB
-
N/A
-
Chat Logs
-
David Crooks: (07/05/2014 11:15)
https://www.gridpp.ac.uk/wiki/Argus_Server#Configuring_Argus_for_Central_Banning
Andrew John Washbrook: (11:22 AM)
monitorama looks good - I like the talk titles
Computers are a Sadness, I am the Cure
James Mickens
David Crooks: (11:23 AM)
Nothing more than normal
Andrew John Washbrook: (11:23 AM)
that could be the new scotgrid tagline
David Crooks: (11:23 AM)
They're streaming on ustream but there will be archives later